Reading Materials

Monetary Intervention Mitigated Banking Panics
during the Great Depression: Quasi-

Experimental Evidence from a Federal Reserve
District Border, 1929–1933
Gary Richardson
University of California, Irvine and National Bureau of Economic Research
William Troost
University of Southern California
The Federal Reserve Act divided Mississippi between the 6th (Atlanta)
and 8th (St. Louis) Districts. During the Great Depression, these dis-
tricts’ policies differed. Atlanta championed monetary activism and
the extension of aid to ailing banks. St. Louis eschewed expansionary
initiatives. During a banking crisis in 1930, Atlanta expedited lending
to banks in need. St. Louis did not. Outcomes differed across districts.
In Atlanta, banks survived at higher rates, lending continued at higher
levels, commerce contracted less, and recovery began earlier. These
patterns indicate that central bank intervention influenced bank
health, credit availability, and business activity.
We thank friends and colleagues for advice and encouragement. Dan Bogart, Michael
Bordo, Jan Brueckner, Charles Calomiris, Michelle Garfinkel, Joseph Mason, Christina
Romer, David Romer, Jean-Laurent Rosenthal, Eugene White, an anonymous referee, and
participants in seminars at the Federal Deposit Insurance Corporation, Federal Reserve
Bank of Atlanta, Federal Reserve Bank of Chicago, Federal Reserve Bank of St. Louis,
Federal Reserve Board of Governors, NBER Summer Institute, Rutgers University, George
Mason University, University of California, Berkeley, University of California, Irvine, and
Western Economics Association provided comments on earlier drafts. This material is based
on work supported by the National Science Foundation under grant 0551232. Shagufta
Ahmed, Shaista Ahmed, Ching-Yi Chung, Joanna Do, and William Troost provided ex-
ceptional research assistance.
[ Journal of Political Economy, 2009, vol. 117, no. 6]

䉷 2009 by The University of Chicago. All rights reserved. 0022-3808/2009/11706-0002$10.00
1031
This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
1032 journal of political economy
I. Introduction
Banks failed throughout the Great Depression. Their demise contrib-

uted to the disruption of financial intermediation that spawned the
deepest downturn in American history (Friedman and Schwartz 1963;
Bernanke 1983; Temin 1989; Romer 1993). The Federal Reserve failed
to stem the falling tide for many reasons. Its leaders adhered to outdated
doctrines and monitored misleading indicators of monetary conditions
(Meltzer 2003, 282, 402–22). The Federal Reserve Board lacked lead-
ership and could not coordinate policies among its disputatious districts
(Wheelock 1991, 72–74, 113–17; Meltzer 2003, 266, 408). The gold
standard fettered mechanisms of monetary policy and the minds of
central bankers (Eichengreen 1992).
Even if the Federal Reserve had tried to alleviate the banking crisis,
no clear evidence exists that it could have helped depository institutions.
Two schools of thought exist on this issue. One school believes that the
principal causes of the banking crises were withdrawals of deposits, il-
liquidity of assets, and the Federal Reserve’s reluctance to act. The Fed
could have alleviated the crisis by acting as a lender of last resort (Fried-
man and Schwartz 1963; Wicker 1996). The second school concludes
that banks failed because the economy contracted. Asset prices fell. Loan
default rates rose. Banks became insolvent. In such circumstances, the
Fed could do little to aid banks by injecting liquidity into the financial
system (Temin 1976; Calomiris and Mason 2003).
These opposing views coexist for several reasons. One is methodo-
logical. None of the studies directly measures the effects of monetary
policy. All infer the Federal Reserve’s abilities indirectly, by interpret-
ing correlations between bank failures, bank characteristics, and the
business cycle. A second is differences in data sources. Friedman and
Schwartz (1963) analyze data on bank suspensions aggregated at the
national level. Their successors scrutinize similar series at lower levels
of aggregation, disaggregated data consisting of samples of national
banks, or panels of banks from within individual cities, states, or Federal
Reserve districts. The most recent and comprehensive work analyzes a
panel of data for all Federal Reserve member banks (Calomiris and
Mason 2003). Future research, Calomiris and Mason indicate (1639),
should analyze data on all banks, multiple measures of financial distress
(such as suspensions and liquidations), and multiple channels of con-
tagion (such as bank runs and correspondent linkages).
Even with such data, analyzing the impact of Federal Reserve policies
would be difficult. At the national level, Federal Reserve policies were
endogenous reactions to ongoing economic events. Changes in Federal
Reserve policy often coincided with changes in fiscal, tariff, and regu-
latory policies and with shocks to the economy for which data are in-

monetary intervention 1033
sufficient or nonexistent. At the district level, the boundaries of Federal
Reserve districts coincided in most cases with state borders. States
changed policies throughout the depression, often at the same time
and occasionally in reaction to actions of the Federal Reserve. Economic
shocks also differed across time and space. The endogeneity of policies,
simultaneous changes in multiple policy dimensions, and the spectrum
of unobserved shocks impede efforts to attribute differences in out-
comes to differences in policies. When observed, correlations between
outcomes and policies might have been caused by phenomena for which
investigators cannot control.
In such circumstances, quasi-experimental econometric strategies
have become increasingly popular. The task is to find a group of banks
that operated in a single regulatory and economic environment but
were exposed to different Federal Reserve policy regimes. Comparing
outcomes across regimes yields insights free from problems of inference
inherent in traditional analysis. The obvious place to seek such a group
is along the Federal Reserve district borders. Borders occasionally di-
vided states. Mississippi is an example. Its northern half lay within the
8th Federal Reserve District (St. Louis). Its southern half lay within the
6th Federal Reserve District (Atlanta). The two districts’ policies differed
dramatically early in the depression. St. Louis was a staunch advocate
of nonintervention. Atlanta was a leading advocate of assisting banks in
need. The St. Louis and Atlanta Feds applied their different policies to
the portions of Mississippi lying within their jurisdictions. The adoption
of these policies preceded the onset of the depression and had little to
do with circumstances in Mississippi, which was a small and peripheral
portion of each Federal Reserve district, and much to do with the phi-
losophies and experiences of the leadership of the two banks. Thus, the
application of Federal Reserve policies to Mississippi possessed the char-
acteristics of an exogenous policy experiment.
This essay analyzes the impact of Federal Reserve policies in the Mis-
sissippi case. Section II describes the data that we analyze. Section III
examines the historical and economic justification for employing quasi-
experimental methods. Section IV examines the link between monetary
policies and banking activity. The analysis progresses through several
stages. The first is a nonparametric examination of the building blocks
of duration analysis: survival and hazard functions. The second is a
parametric analysis of our panel of data. Our results indicate that ag-
gressive discount lending would have helped banks survive the initial
banking panic of the Great Depression in the fall of 1930. Section V
examines the robustness of this result. Section VI examines the link
between bank failures, credit contraction, and commercial activity. A
clear correlation exists between bank failures during the panic of 1930
and subsequent declines in wholesale trade. Bank distress at other times

TABLE 1
Number of Banks in Mississippi on July 1 of Each Year
State Charter National Charter

Federal Reserve District Federal Reserve District
Year All 6th Atlanta 8th St. Louis All 6th Atlanta 8th St. Louis
1929 274 120 155 35 21 14
1930 259 105 154 35 22 13
1931 222 96 126 28 18 10
1932 206 89 108 27 18 9
1933 189 82 106 24 15 9
Source.—Rand McNally Bankers’ Directory, various July issues, 1929–35.
(e.g., during the fall of 1931) and for other reasons (e.g., loan defaults)
shows little (or no) correlation with subsequent commercial contraction.
Our methods directly address key questions concerning the collapse
of the banking system during the early 1930s. Did Federal Reserve pol-
icies influence bank failure rates? Did providing liquidity (or credibly
committing to do so) reduce rates of bank suspension and liquidation?
Did the demise of banks reduce the supply of commercial credit? Did
the contraction of credit reduce commercial activity? To each of these
questions, the answer is yes.
Section VII discusses the implications of our analysis. By injecting
liquidity into the banking system, particularly during the banking panic
in the fall of 1930, the Federal Reserve Bank of Atlanta prevented bank
failures rates and facilitated business activity. If other Federal Reserve
banks had pursued similar strategies, fewer banks would have failed,
and the depression may have followed a different course.
II. Data Sources

A spectrum of sources provides the essential evidence. The Rand McNally
Bankers’ Directory describes bank balance sheets, correspondent relation-
ships, and characteristics. Observations drawn from July issues provide
a panel of annual observations on state and national banks in Mississippi
at their spring calls. Tables 1 and 2 recapitulate this information. The
U.S. censuses of agriculture, manufacturing, and population describe
the characteristics of counties. Summary statistics appear in table 3.1
Bradstreet’s Weekly, Dun’s Review, the Commercial and Financial Chronicle,
the Federal Reserve Bulletin, and the annual reports of the Federal Reserve
Board and the Federal Reserve banks provide information on building
1
These data can be downloaded from the Interuniversity Consortium for Political and
Social Research, Study 3, Historical, Demographic, Economic, and Social Data: The United
States, 1790–1970 (http://www.icpsr.umich.edu/icpsrweb/ICPSR/.

TABLE 2
Characteristics of Banks in Mississippi on July 1, 1929
6th Federal Reserve District (Atlanta) 8th Federal Reserve District (St. Louis)
All 6th (N p 141) Near Border (N p 76) Near Border (N p 169) All 8th N p 112)
Standard Standard Standard Standard
Median Mean Deviation Median Mean Deviation Median Mean Deviation Median Mean Deviation
Financial ratios:
Net worth/total assets .10 .11 .04 .10 .11 .04 .13 .14 .06 .11 .13 .05
Cash/total assets .37 .38 .14 .36 .39 .14 .38 .37 .15 .38 .38 .15
Deposits/total liabilities .87 .85 .07 .88 .85 .08 .85 .82 .11 .86 .83 .10
Financial characteristics:
Total assets ($1,000) 559 1,166 141 514 1,211 225 451 790 106 448 748 76
Loans and discounts ($1,000) 334 676 1,070 278 713 1,288 270 464 755 256 437 668
Cash and exchanges ($1,000) 92 204 310 84 228 373 92 174 276 91 157 237
Deposits ($1,000) 506 1,003 1,445 465 1,040 1,699 379 662 993 369 629 869
Paid-up capital ($1,000) 30 59 75 30 63 86 30 52 65 30 49 57
State-chartered banks (%) .85 .36 .88 .33 .90 .30 .92 .28
Federal Reserve member (%) .15 .36 .12 .33 .12 .32 .10 .30
Years in operation 24 23.2 12.3 24.5 24.0 12.7 21 21.9 14.9 20.5 21.8 14.2
Correspondents (N) 3 3.10 .90 3 3.08 .95 3 3.04 .89 3 2.96 .96
Source.—Rand McNally Bankers’ Directory, various July issues, 1929–35.
Note.—Near border sample consists of banks in counties for which at least 50 percent of the area lies within 1 degree latitude of the Federal Reserve district border.

TABLE 3
Characteristics of Counties in Mississippi in 1930
8th Federal Reserve District

6th Federal Reserve District (Atlanta) (St. Louis)
All Near Border Near Border All
Standard Standard Standard Standard
Mean Deviation Mean Deviation Mean Deviation Mean Deviation
Population (1,000s) 22.4 14.4 28.2 17.7 30.4 17.2 26.8 14.2
Persons per square mile 37.4 19.7 41.5 20.3 51.4 21.5 49.3 18.6
Urban population share (%) 14.2 22.3 12.2 22.8 12.5 11.1 9.3 10.8
Black population share (%) 43.4 18.2 49.5 18.2 56.1 18.1 49.6 23.3
Number of manufacturing establishments 20.1 20.0 25.6 24.6 27.1 14.1 25.2 15.9
Average annual manufacturing wage ($) 754.8 150.6 779.2 129.3 753.7 182.9 711.2 178.7
Net sales, retail stores, annual per capita
($) 190.0 76.8 188.2 91.7 185.0 51.5 175.1 54.0
Fraction of population in labor force (%) 38.8 6.2 41.3 6.3 42.9 7.6 42.4 8.0
Unemployment rate (%) 1.8 2.0 1.0 1.1 .6 .4 .5 .4
Fraction of farm acres in cotton (%) 57.5 26.4 68.0 18.2 77.7 14.1 79.7 11.9
Fraction of farm acres with crop failures
(%) 3.3 6.4 3.8 7.3 1.1 .5 1.1 .8
Farm mortgage debt as a percentage of
farm value 33.2 5.3 35.3 4.2 41.2 7.2 41.6 6.1
Interest charges as a percentage of mort-
gage debt 7.0 .5 6.9 .4 6.9 .5 6.9 .4
Source.—Historical, Demographic, Economic, and Social Data: The United States, 1790–1970 (http://www.icpsr.umich.edu/icpsrweb/ICPSR/. For comparisons of additional characteristics,
see Richardson and Troost (2006).
Note.—The near border columns contain statistics for counties for which at least 50 percent of the area lies within 1 degree latitude of the Federal Reserve district border.

permits, business failures, commodity prices, interest rates, and price
and production indices.
The archives of the Federal Reserve Board provide additional infor-
mation. Forms St. 6386a and 6386c from the Division of Bank Opera-
tions report changes in bank status including mergers and reopenings
of suspended banks. Form St. 6386b reports bank suspensions and their
causes. The reports distinguish between temporary and permanent sus-
pensions. A temporary suspension occurred when a bank closed its doors
to depositors for at least one business day and later resumed operations.
Permanent suspensions, which we refer to as liquidations, were the subset
of suspensions in which insolvent banks ceased operations, surrendered
charters, and repaid creditors under the auspices of a court-appointed
receiver.2
The U.S. Census of American Business (U.S. Census Bureau 1935) pro-
vides information on commercial activity. The census reports the num-
ber of establishments engaged in wholesale distribution and net sales
of those establishments in 1929 and 1933. The figures for 1929 indicate
the amount of wholesale activity in each county in the peak year prior
to the contraction. The figures for 1933 indicate the amount of whole-
sale activity in each county at the trough of the depression.
These sources provide a panel of data consisting of all banks that
operated in Mississippi between July 1929 and July 1933. The panel
contains information about bank characteristics and economic condi-
tions, which other scholars have used, and information about multiple
measures of financial distress (such as suspensions and liquidations),
other changes in bank status (such as mergers, consolidations forced
by financial difficulties, and voluntary liquidations), multiple paths of
contagion (including correspondent linkages and runs on banks), fac-
tors fundamental to the performance of the national economy and
particularly pertinent to Mississippi (such as levels of farm indebtedness
and the condition of the cotton crop), and measures of Federal Reserve
policy regimes, which scholars have not employed in the past.
A wide variety of historical sources reveal the policy regimes of the
Atlanta and St. Louis Federal Reserve banks. The archives of the Board
of Governors contain correspondence between the Board, the Atlanta
Fed, and the St. Louis Fed that describes their actions and illuminates
their intentions. So do articles in depression-era newspapers and Gam-
ble’s (1989) in-house history of the Atlanta Fed. The annual reports of
the Reserve banks also describe their policies and the implementation
of their plans.
2
These records reside in Record Group 82, “Central Subject File of the Federal Reserve
Board of Governors, 1913–1954,” National Archives and Records Administration, College
Park, MD. For detailed descriptions of these data, see Richardson (2006, 2007a, 2007b,
2008).

Four independent sources enable us to determine the dates and the
nature of Mississippi’s banking crises. The first is the Federal Reserve’s
St. 6386 database, which indicates the dates and causes of bank sus-
pensions. The second is a narrative description of events contained
within the biennial reports of Mississippi’s state Banking Department.
The third is articles in an array of newspapers including three with the
largest circulation in Mississippi, the Meridian Star, Vicksburg Herald, and
Vicksburg Sunday Post-Herald; the leading papers from the headquarters’
cities of the 6th and 8th Federal Reserve Districts, the Atlanta Journal,
St. Louis Globe-Democrat, and St. Louis Post-Dispatch; and the New York Times.
The fourth is a Department of Agriculture study into the seasonal pat-
tern of bank balances in Mississippi from 1923 through 1936 (Wall
1937).
III. Historical Background

Our quasi-experimental approach builds on three facts. First, when the
depression began, the policy regimes of the Atlanta and St. Louis Federal
Reserve banks differed, and those differences were exogenous to the
state of Mississippi and events occurring at the time. Second, Mississippi
was homogeneous in economic, demographic, and regulatory dimen-
sions, particularly in counties adjacent to the Federal Reserve district
boundary. Third, in the fall of 1930, Mississippi experienced a panic in
which a sudden shift in depositors’ perceptions about the safety of fi-
nancial institutions triggered runs on banks, and depositors withdrew
funds en masse
A. Policy Regimes
Friedman and Schwartz (1963) pioneered efforts to identify Federal
Reserve policy regimes using the narrative historical approach. Romer
and Romer (1989) emphasize the importance of establishing clear cri-
teria for identifying policy regimes. Since our essay focuses on banking
crises, we define regimes in terms of policies regarding whether and
how to intervene during widespread and rapid failures of financial in-
stitutions that contemporaries classified as panics.
In the spring of 1913, the organizing committee of the Federal Re-
serve System split the state of Mississippi evenly between the 6th and
8th Districts. Figure 1 depicts the division (as well as information dis-
cussed later). In the 6th District from 1913 until the mid-1930s, the
Atlanta Fed followed Bagehot’s rule, a doctrine that during financial
panics, central banks should act as lenders of last resort and extend
credit to institutions afflicted by illiquidity. Such lending should be suf-
ficient to enable solvent but illiquid institutions to survive deposit losses

Fig. 1.—Mississippi’s division into Federal Reserve districts and bank suspensions be-
tween October 1930 and March 1931. Source: See Section II. The solid line represents
the Federal Reserve district border. The dotted lines enclose the counties for which at
least half the area lies within 1 degree latitude of the district border.
and, thus, to prevent runs from driving healthy banks into insolvency.
From 1913 to 1929, the Atlanta Fed faced four panics in which it could
employ such policies. In each instance, the Atlanta Fed rushed large
quantities of cash to the afflicted region, extended emergency loans to
member banks, helped member banks extend credit to their country
clients, and returned the situation to status quo ex ante (for details, see
Richardson and Troost [2006]). During the depression of the 1930s,
the Atlanta Fed consistently advocated monetary expansion. Atlanta’s
advocacy caught President Franklin Roosevelt’s attention, and he ap-
pointed Atlanta’s governor, Eugene Black, to be the chairman of the
Federal Reserve Board.
The policies of the Federal Reserve Bank of St. Louis were far dif-

ferent. The St. Louis Fed was a principal proponent of the “Real Bills
view that the supply of credit should contract during recessions” since
a lower level of economic activity required less credit to sustain it (Whee-
lock 1991, 53, 111). According to this doctrine, excessive credit expan-
sion generated fears of inflation and uncertainty about interest rates,
which deterred investment and retarded recovery. For this reason, the
directors “opposed reductions in discount rates and other actions
[which would] retard the necessary process of liquidation” (Chandler
1971, 142). The 8th District held this view because the city of St. Louis
served as an agricultural entrepot, and the Federal Reserve Bank of St.
Louis focused on accommodating the strong seasonal business cycle by
procyclically expanding and contracting credit.
Because the St. Louis Fed adhered to the doctrine of real bills, it ran
a tight discount window. During periods of panic, it limited lending and
at times refused requests to rediscount eligible paper. When it did ex-
tend loans, the St. Louis Fed required what was then known as marginal
or double collateral. This collateral consisted of the eligible paper re-
quired by law plus an equal amount of U.S. government securities. This
practice discouraged banks from using the discount window as a source
of liquidity, since they had to turn over $2 of their most liquid assets
to get $1 of cash (Westerfield 1932). The St. Louis Fed maintained this
policy until the summer of 1931. In July of that year, the St. Louis Fed
ceased opposing intervention, eased collateral requirements, and ex-
panded lending through the discount window (Chandler 1971, 142).
B. Homogeneous Conditions
Mississippi was homogeneous economically and demographically. Table
3 demonstrates this by displaying county-level data drawn from the cen-
suses of population, manufacturing, and agriculture for 1930. In both
districts, the fraction of the population in the labor force was substantial.
Unemployment rates were low. Farm debt hovered around one-third to
one-fifth of farm value. Rural counties concentrated on cultivating cot-
ton. Prevailing prices for labor (average annual manufacturing wage)
and capital (ratio of interest charges to mortgage debt) differed little
across counties. The largest differences arose in the extremities of the
state. The counties adjoining the Federal Reserve district border had
few discernible differences.
Mississippi’s banking system was also homogeneous. The Banking De-
partment applied standard procedures throughout the state. So did the
Office of the Comptroller of Currency, Reconstruction Finance Cor-
poration, Department of Agriculture, Works Progress Administration,
and the intermediate credit banks, since Mississippi lay within a single
district for all these federal bureaus.

Mississippi’s unit banking system resembled that prevailing through-
out the United States. The law prohibited branching. Banks operated
out of a single building. They served areas near their headquarters, with
most depositors and borrowers belonging to the local community. Few
local banks belonged to the Federal Reserve System. Instead, they relied
on correspondents in municipalities designated reserve cities, which be-
longed to the Federal Reserve System. Correspondent banks provided
local banks with liquidity, often by discounting short-term commercial
paper, which correspondents then rediscounted at a Federal Reserve
facility. This financial structure enabled the Federal Reserve to influence
the liquidity of nonmember institutions by encouraging correspondents
to extend credit (or discouraging them from withholding credit).
C. The Banking Crisis

On November 7, 1930, Caldwell and Company collapsed in Nashville,
Tennessee. Caldwell controlled the largest financial conglomerate in
the South, and its principal affiliate, the Bank of Tennessee, served as
the correspondent for hundreds of financial institutions. During the
next few days, correspondent networks toppled like dominoes in Ten-
nessee, Arkansas, Illinois, and North Carolina. Bank runs radiated geo-
graphically from the locus of the counterparty cascade. Caldwell’s cor-
respondent network did not extend into Mississippi, where the banking
situation remained stable for 6 weeks (Richardson 2007b).
During this period, newspapers in Mississippi reported the financial
scandal underlying Caldwell’s demise (e.g., Vicksburg Herald, November
8, 1930, 1). Newspapers also reported on the financial scandals con-
tributing to the closure of the Bank of the United States in New York
City (Atlanta Journal, December 11, 1930, 33; December 12, 1930, 36;
December 16, 1930, 29) and the closure of the Guaranty Building and
Loan Association in Hollywood, California (Atlanta Journal, December
12, 1930, 1, 10). Newspapers emphasized a court decision that invali-
dated a law that exempted Mississippi banks from taxation (Meridian
Star, December 1, 1930, 1). The decision threatened to increase banks’
operating expenses and weaken their financial positions. The decision
also cast doubt on Mississippi’s recently revised banking codes and
threatened to saddle banks with large liabilities from the deposit in-
surance program, which the state discontinued in the spring.
The incessant discussion of financial corruption, banking panics, in-
dustrial recession, and court cases appears to have taken a toll on de-
positors’ confidence. The Vicksburg Herald’s weekly tabulation of Vicks-
burg bank balance sheets shows deposits falling at an increasing rate
during November and December. The process remained orderly until
Friday, December 19, 1930, when panic struck. On that day, the state

Banking Department closed three banks: one because of embezzlement
and two because of frozen assets and poor collections. The next day,
one of the larger banks in the state placed itself in the hands of the
Banking Department after the suspicious death, rumored to be a suicide,
of its president.
Rumors triggered runs on nearby banks, which soon spread to neigh-
boring towns, and within a week, throughout the state. Forty-nine in-
stitutions quickly closed their doors to depositors. The law allowed banks
to suspend operations for up to 5 days at directors’ discretion and for
a longer period if they could demonstrate both compelling necessity
and the ability to reopen after the crisis passed. Banks that remained
in operation slowed withdrawals from savings accounts by requiring 30
days’ notice for withdrawals (a provision in most deposit contracts) and
refusing to terminate time deposits before maturity.
During this whirlwind of withdrawals, aggregate deposits at Missis-
sippi’s banks fell by nearly 30 percent, even though four out of 10 banks
closed their doors to depositors and the remainder slowed withdrawals
however possible. Lending by banks fell by a similar percentage (Biennial
Report of the Banking Department of the State of Mississippi, 1931, 4). In prior
Decembers and Januaries, balances had increased in Mississippi’s banks,
as farmers deposited proceeds from the harvest. The typical end-of-year
inflow increased deposits from 20 percent to 30 percent. Thus, at a time
when Mississippi’s banks expected large inflows of funds, they experi-
enced large outflows. When withdrawals abated in February 1931, the
level of deposits was 55 percent lower than a year earlier and 65 percent
lower than the average of the preceding five Februaries. Deposits and
lending remained near this nadir until 1934 (Wall 1937, 20).
D. The Historical Experiment
Responses to this banking panic create the historical experiment. The

6th District reacted as it had in the past. It rushed cash in large quantities
to banks undergoing runs. It extended credit to member banks as
quickly and substantially as possible and encouraged them to extend
loans to their counterparties. During the three weeks following Cald-
well’s collapse, discounts to member banks increased by $2,800,000, and
total Federal Reserve credit to member banks increased by more than
$8,100,000 (Wicker 1996, 54).
The 8th District also acted as it had in the past. The St. Louis Fed
made no effort to expedite discount lending and may have slowed dis-
bursement by more stringently monitoring the quality of paper sub-
mitted for rediscounting. During the first three weeks of the crisis, dis-
counts to member banks in the 8th District declined by $2,100,000, and

Fig. 2.—Discount response after the collapse of Caldwell, aggregate discounts each week
as a percentage of initial level. Source: See Section II.
total Federal Reserve credit to member banks in the 8th District declined
by more than $11,800,000 (Wicker 1996, 54).
Figure 2 illuminates changes in discount lending following Caldwell’s
collapse. Discounts of the 6th District rose rapidly to a peak 40 percent
higher than before the crisis. Discounts of the 8th District fell gradually,
as the extension of new loans slowed and existing loans expired.
In this situation, models of the Diamond-Dybvig type provide clear
predictions (Diamond and Dybvig 1983). Bank failure rates in the 6th
District should have been lower than bank failure rates in the 8th Dis-
trict, since a lender of last resort can mitigate financial panics by ex-
tending credit to illiquid institutions (and perhaps forestall a panic by
credibly committing to do so).3 Since the Atlanta Fed implemented such
a policy in a prompt, ample, and public manner, difference in outcomes
between the 6th and 8th Districts (if any) should reveal the effectiveness
(or ineffectiveness) of Atlanta’s policies.
Nonpanic periods serve as a control case that helps to test the ho-
mogeneity assumption underlying our analysis. Bagehot’s rule is a policy
implemented during panics, when withdrawals, contagion, and illiquid-
ity bedevil banks. The policy does not operate and therefore should
have no direct effect on bank failure rates outside of panic periods. The
period following the summer of 1931, when the St. Louis Fed adopted
3
Recent theoretical work indicates that monetary expansion can prevent the failure of
depository institutions, curtail contagion among banks, and alleviate the contraction of
liquidity even when real shocks (such as loan defaults or delays in bond repayment) are
the root of the problem and trigger the illiquidity crisis. See Diamond and Rajan (2005,
2006).

TABLE 4
Bank Suspensions and Liquidations
Percentage of Banks Percentage of Banks

Suspending Liquidating
Federal Reserve District Federal Reserve District
Begin End All 6th Atlanta 8th St. Louis All 6th Atlanta 8th St. Louis
July 1 June 30 (1) (2) (3) (4) (5) (6)
1929 to 1930 4.8 7.1 3.0 4.5 7.1 2.4
1930 to 1931 28.9 14.2 39.5 13.6 7.1 18.6
1931 to 1932 13.2 14.9 11.8 8.0 7.9 8.1
1932 to 1933 7.7 7.5 7.9 7.3 6.5 7.9
1933 to 1934 .9 .0 1.7 .9 .0 1.7
1929 to 1934a 49.8 38.7 59.2 30.9 26.8 34.4
Source.—Rand McNally Bankers Directory and National Archives and Records Administration Record Group 82. See
Section II and Richardson (2006, 2007a, 2007b, 2008) for details.
a
The last row indicates the percentage of banks operating on July 1, 1929, that either suspended or liquidated by
June 30, 1933.
policies resembling those in Atlanta, also serves as a control case. If

policies mattered, differences in outcomes should exist when policies
differed but should not occur when policies were the same.
E. Patterns Apparent in Basic Data

Table 4, figure 1, and figure 3 illuminate the outcomes of interest. Table
4 reports suspension and liquidation rates for each year from July 1929
to July 1934. The rates peaked in the second year of the depression and
remained above predepression levels until the national banking holiday
in March 1933. Table 4 shows that when the Atlanta and St. Louis Feds
pursued opposite policies during the fall and winter of 1930, fewer banks
failed in the 6th District, which made every effort to inject liquidity into
the banking system. More banks failed in the 8th District, which
preached nonintervention and where Federal Reserve credit outstand-
ing fell substantially. Afterward, as the policies of the districts converged
and the nature of the banking difficulties changed, rates of suspension
and liquidations also converged.
Figure 3 illustrates these patterns by plotting the percentage of “banks
in operation” and “banks in business” each day over the entire span of
our data panel from July 1, 1929, to June 30, 1933.4 Banks in operation
4
The numerator of the series banks in operation is the number of banks in operation
on July 1, 1929, minus the number of banks that since that date suspended operations
(either temporarily or permanently), consolidated because of financial distress, liquidated
voluntarily, or surrendered their charter after merging with another institution and plus
the number of banks that since July 1, 1929, newly opened for business or reopened after
temporarily suspending operations. The numerator of the series banks in business equals
banks in operation plus the number of temporarily suspended banks. The denominator
of both series is the number of banks in operation (which equals the number of banks
in business) on July 1, 1929.

Fig. 3.—Percentage of banks in business and in operations in the 6th and 8th Federal
Reserve Districts in Mississippi, July 1929 to June 1933. Source: See Section II.
are banks whose doors are open to the public. Banks in business are
banks that are not bankrupt. The difference is the number of tempo-
rarily suspended banks. Figure 3 also indicates the date when the St.
Louis Fed’s policies began to converge toward those of the Atlanta Fed
and the dates of the events that the historical literature identifies as
triggers of the surges in suspensions apparent in the evidence. Figure
3 shows that during the post-Caldwell panic, when policy regimes dif-
fered across districts, banks suspended operations (temporarily and per-
manently) at much higher rates in the 8th District. During later surges
in bank suspensions, when policies differed little and the rise in failures
stemmed largely from fundamental factors, banks in the 6th and 8th
Districts failed at similar rates.
Figure 1 plots on a map of Mississippi each bank that suspends from
October 1930 through March 1931. The map indicates the division of
the state into the 6th (below solid line) and 8th (above solid line)
Districts. The area between the dotted lines indicates counties for which
at least half of the area lies within 1 degree latitude of the Federal
Reserve district border. The events precipitating the panic began in the
town that had five bank suspensions. The suspensions appear to be
distributed widely throughout both districts without prominent geo-
graphic patterns.

IV. Methods and Results
Statistical analysis substantiates these observations by controlling for
characteristics of individual banks, the economic environment, and
other phenomena that might have generated the observed differences
across districts. We first control for potentially confounding factors non-
parametrically. Then we present parametric estimates.
A. Nonparametric Estimates
The analysis of time to failure rests on survivor and hazard functions.
This subsection presents nonparametric estimates of survivor functions
constructed via the Kaplan-Meier method and of hazard functions con-
structed by smoothing raw hazard rates (i.e., the number of bank failures
divided by the number of banks at risk on each date). Kernels are
discrete Epanechnikov. Bandwidths of 28 days on graphs spanning 4
years and 7 days on graphs spanning 4 months are wide enough to
smooth daily volatility without obscuring weekly shifts in the probability
of failure.5
Figure 4 presents survival and hazard functions for all banks in Mis-
sissippi during the banking crisis in the fall of 1930. The time under
analysis is restricted to the four months following the collapse of Cald-
well and Company. The population at risk is all banks in operation. A
bank that surrendered its charter voluntarily or merged with another
institution departs from the population at risk (but is not counted as a
failure) on the date that it ceased operations. A bank that suspended
operations is counted as a failure on the date that it closed its doors to
the public.
5
Our estimates of the survival function, S(t) , the raw hazard function, h(t), and the
smoothed hazard function, g(t), are
Ŝ(t) p
ti!t
写
ni ⫺ di
ni
,
where ni is the number of banks in business at the beginning of time period ti , di is the
number of banks experiencing an event (such as entering receivership) at time ti , and
ti indicates the ith time period. The raw hazard for period ti is
di
ĥ(ti) p .
ni
The hazard function is estimated by smoothing raw hazards, so that the hazard in the ith
time period is
冘
u
ĝ(ti) p ˆ
Kzh(ti⫹z),
zp⫺u
where u is the bandwidth and

(u ⫹ 1)2 ⫺ z2
Kz p
冘
zp⫺u
u
[(u ⫹ 1)2 ⫺ z2]
.

In figure 4, the gray lines depict the 6th District. The dotted lines
depict the 8th District. Figures 4A and B show that following Caldwell’s
collapse, patterns of hazard and survival differed dramatically between
the 6th and 8th Districts. Failure rates in the 8th District rose rapidly
and exceeded those in the 6th District for most of the crisis. The array
of nonparametric tests for the equality of survival functions—including
the log rank, Breslow, Peto-Peto, and Tarone-Ware tests—reject at the
1 percent significance level the null hypothesis that the survival function
for the 6th District equaled that for the 8th District. All the tests produce
x2 statistics (with one degree of freedom) of over 20.
The remainder of figure 4 demonstrates that differences in suspension
rates across districts during the post-Caldwell panic cannot be attributed
to fundamentals or selection. Figures 4C and D limit the analysis to
banks that operated within 1 degree latitude of the Federal Reserve
district border.6 These figures demonstrate that even in a narrow band
along the border, banks failed at a higher rate in the 8th District and
a lower rate in the 6th District. Economic fundamentals varied little
over such short distances, particularly in economically and politically
homogeneous central Mississippi. Thus, differences in fundamentals
were not the reason that failure rates differed between districts.
Figures 4E and F limit the analysis to banks in operation before the
founding of the Federal Reserve in 1913. Figures 4G and H limit the
analysis to banks founded after the Federal Reserve System. These fig-
ures demonstrate that banks in both groups failed at a higher rate in
the 8th District. Therefore, selective pressures, which would have altered
the pattern for these groups, were not the reason that failure rates
differed between districts.
Figure 5 illustrates patterns of suspensions over the entire sample
period. The event under analysis is suspension of operations. The def-
inition of the population at risk remains as above except for temporarily
suspended banks, which depart the population at risk when suspended
and reenter the population at risk after resuming operations. All the
graphs depict a similar pattern. In the 8th District, more banks failed,
and failures were clustered during periods of panic. In the 6th District,
fewer banks failed, particularly during the banking panic of 1930, and
failures were spaced more evenly through time.
Figures 5A and B illuminate important issues. During nonpanic pe-
riods, the suspension rate in the 6th District exceeded that of the 8th
6
Throughout this essay, whenever we state “within 1 degree latitude of the border,” we
are referring to this county-based distance definition. The set includes all banks operating
in a county for which at least 50 percent of the surface area lay within 1 degree latitude
of the border. This definition generates a band running through the center of the state
straddling the Federal Reserve district border. The outer edges of the band vary from 70
to 95 miles distance from the boundary.

Fig. 4.—Survival and hazard during the post-Caldwell panic, principle nonparametric
controls. A, Survival within 1 degree latitude of border. C, Survival within 50 miles of
border. E, Survival for banks founded before Fed. G, Survival for banks founded after Fed.
1048

Fig. 4.—Survival and hazard during the post-Caldwell panic, principle nonparametric
controls (continued). B, Hazard within 1 degree latitude of border. D, Hazard within 50
miles of border. F, Hazard for banks founded before Fed. H, Hazard for banks founded
after Fed.
1049

Fig. 5.—Bank suspension in the 6th and 8th Federal Reserve Districts, July 1929 through
February 1933. A, Survival for all banks. C, Survival within 1 degree latitude of border. E,
Survival within 50 miles of border. G, Survival for banks founded before Fed.
1050

Fig. 5.—Bank suspension in the 6th and 8th Federal Reserve Districts, July 1929 through
February 1933 (continued). B, Hazard for all banks. D, Hazard within 1 degree latitude of
border. F, Hazard within 50 miles of border. H, Hazard for banks founded before Fed.
1051

District, particularly in the period preceding the collapse of Caldwell,
when principal employers in two towns in the southern half of the state
closed, forcing nearby banks out of business. This pattern suggests that
economic fundamentals favored banks in the 8th District over those in
the 6th District. During periods of panic, however, banks in the 8th
District failed at higher rates. This pattern is consistent with the effective
application of Bagehot’s rule, which should reduce failure rates during
panics, when the lender of last resort loans freely, but not during normal
times, when the lender of last resort husbands its reserves and allows
insolvent banks to liquidate.
The remaining figures demonstrate the robustness of the result. Fig-
ures 5C and D limit the analysis to all banks that operated within 1
degree latitude of the border. Figures 5E and F limit the analysis to all
banks that operated within 50 miles of the border. Figures 5G and H
limit the analysis to banks established before 1913. In each case, the
pattern remains the same.
The pattern also remains the same when we limit analysis to groups
of banks with similar characteristics, such as longevity or stable man-
agement, or groups of banks operating in similar environments, such
as cities or cotton-growing regions. Subpopulations that we have ex-
amined include state-chartered banks, nationally chartered banks, Fed-
eral Reserve member banks, Fed nonmember banks, banks in the west-
ern and eastern halves of the state, banks in operation for more or fewer
years than the median age of all banks, banks with and without man-
agement changes between 1925 and 1929, banks in counties with more
and less than the median percentage of agricultural acreage dedicated
to cotton cultivation, and banks in counties with above and below the
median number of manufacturing establishments. When the measure
of distress is changed to liquidation, the interdistrict differences in bank
failure rates retain the same sign but increase in magnitude. The in-
variance of the pattern across measures of distress and across subpop-
ulations defined by likely correlates with economic fundamentals and
selected characteristics suggests that neither fundamentals nor selection
drives our results.
We confirm the differences apparent in the pictures with the appro-
priate nonparametric tests for the equality of survivor functions. In all
cases, the null hypothesis of the equality of the survival functions in the
6th and 8th Districts can be rejected for the post-Caldwell panic. Similar
hypothesis tests for the equality of survival functions following Britain’s
departure from the gold standard in the fall of 1931 and Roosevelt’s
election in the fall of 1932, when gold outflows forced the Federal
Reserve to raise interest rates and bank failures increased throughout
the nation, cannot reject the null hypothesis. This result corroborates
the homogeneity assumption underlying our analysis, that banks in

northern and southern Mississippi operated in similar economic envi-
ronments, faced similar challenges, and experienced similar outcomes,
except during the panic in the fall of 1930, when discount-lending
policies differed between districts.
The tripartite pattern apparent in figure 5—(1) hazard rates for the
6th and 8th Districts similar at all times except during the panic fol-
lowing Caldwell’s collapse, when the hazard for the 8th District exceeded
that in the 6th by a wide margin; (2) cumulative hazard for the entire
period higher in the 8th District; and (3) failures clustered during pe-
riods of heightened risk—appears robust to alterations in our non-
parametric framework. A nonparametric test for this pattern, however,
does not exist. Generating such tests requires additional assumptions.
For this task, we turn to parametric methods.
B. Parametric Estimates
A plethora of potential parameterizations exist for our analysis. We pre-
sent results for the current gold standard in this literature, the log-
logistic survival model of Calomiris and Mason (2003). In this model,
the unit of observation is the individual bank. The dependent variable
is log days until distress. Time under observation begins on July 1, 1929,
and ends at the national banking holiday in March 1933. The explan-
atory variables include the characteristics of banks, the characteristics
of counties in which banks operate, a measure of business conditions
at the national level, indicators of periods of panic, and in our version
of this model, indicators of Federal Reserve policy regimes. Bank char-
acteristics update annually each July 1. County characteristics (from the
Census of 1930) remain constant over time. Economic conditions up-
date monthly. This framework allows us to determine the relative im-
portance of fundamentals and contagion as sources of bank distress and
to test whether Federal Reserve intervention mitigated (or accentuated)
banking panics.
Table 5 presents the results of this exercise. Column 1 reports the
basic model. It contains indicator variables for the three surges in bank
suspension in the fall of 1930, fall of 1931, and winter of 1933; for
whether a bank operated within the 6th District; and for whether during
each of three surges in bank suspensions a bank operated within the
6th District. The crisis indicators reveal to what extent failure rates rose
above the baseline during each surge. The crisis/district interaction
terms reveal for each crisis whether failure rates differed between the
6th and 8th Districts. The coefficients for the banking crises in 1930
and 1931 are statistically significant, indicating that during the crises,
the rate of bank distress rose above the baseline. The coefficient for
the fall 1930 crisis/Atlanta Fed interaction term is also statistically sig-

TABLE 5
Log-Logistic Survival Regressions for Individual Banks
Dependent Variable: Log Days until Bank Distress
(1) (2) (3) (4) (5) (6)

Fed Atlanta during crisis 1930 11.20* 2.65* 2.45* 2.64* 2.82* 1.96*
1054
(1.18) (.78) (.67) (.76) (1.35) (.53)
Fed Atlanta during crisis 1931 1.25 .42 .47 .31 .13 .28
(.84) (.65) (.59) (.64) (1.23) (.50)
Fed Atlanta during crisis 1933 .61 1.07 .97 .86 .73 .77
(.92) (.94) (.90) (.98) (1.35) (.72)
Federal Reserve Atlanta ⫺1.01* ⫺.90* ⫺.98* ⫺1.05* ⫺.93** ⫺.77*
(.38) (.36) (.47) (.43) (.52) (.33)
Banking crisis—fall 1930 ⫺12.38* ⫺3.76* ⫺3.50* ⫺3.70* ⫺4.19* ⫺2.67*
(1.19) (.92) (.77) (.92) (1.60) (.58)
Banking crisis—fall 1931 ⫺2.85* ⫺2.41* ⫺2.23* ⫺2.34* ⫺2.66* ⫺1.88*
(.74) (.59) (.54) (.56) (.78) (.45)
Banking crisis—winter 1933 ⫺1.00 ⫺1.51** ⫺1.40** ⫺1.50** ⫺1.46 ⫺.47
(.70) (.87) (.81) (.87) (1.33) (.56)
Assets % cash 6.37* 5.66* 6.31* 7.06* 4.55*
(1.51) (1.41) (1.53) (2.35) (1.142

Net worth/total assets 7.53* 6.08* 6.60* 6.57* 6.96*
(2.21) (1.94) (2.19) (2.54) (2.29)
Constant 9.09* 5.59* 4.97* 6.99* 11.31* 68.41
(.35) (1.97) (2.09) (2.14) (3.65) (78.73)
ln(g) ⫺.37* ⫺.65* ⫺.70* ⫺.67* ⫺.61* ⫺.82*
(.09) (.11) (.11) (.12) (.13) (.12)
Bank characteristics vector MS MS MS MS CM
County characteristics vector MS PC PC CM
Number of explanatory variables 9 17 21 22 22 28
Sample limited to Border
counties
Number of subjects 318 318 318 318 181 318
Number of failures 129 129 129 129 72 129
Days at risk 314,954 314,954 314,954 314,954 176,581 314,954
Log likelihood ⫺186.3 ⫺144.2 ⫺140.3 ⫺141.8 ⫺78.8 ⫺133.8
Wald x2 115.6 34.3 44.6 40.6 35.4 98.35
Wald x2 degrees of freedom 7 15 19 20 20 26
Note.—The dependent variable is log days until distress (liquidation, suspension, or consolidation under duress) after July 1, 1929. Standard errors (in parentheses) estimated with Huber-
1055
White sandwich method clustered on individual banks. CM indicates that the vector of control variables conforms to the specifications of Calomiris and Mason (2003). MS indicates that the vector
is fitted to Mississippi fundamentals. PC indicates that the vector consists of principal components of county variables. The border county sample consists of all banks located in counties for which
at least 50 percent of the area lay within 1 degree latitude of the Federal Reserve district border. The number of explanatory variables includes the constant and the curvature parameter.
* Significant at the 5 percent level.
** Significant at the 10 percent level.

TABLE 6
Magnitudes of Effects of Policy Regimes and Panics: Change in Cumulative
Hazard Rates in Log-Logistic Regressions
(1) (2) (3) (4) (5) (6)

Fed Atlanta during
panic 1930 ⫺30.5 ⫺32.6 ⫺33.2 ⫺33.2 ⫺27.8 ⫺37.4
Fed Atlanta during
panic 1931 ⫺6.2 ⫺2.6 ⫺3.2 ⫺2.0 ⫺.6 ⫺2.5
Fed Atlanta during
panic 1933 ⫺1.7 ⫺3.1 ⫺3.0 ⫺2.5 ⫺1.4 ⫺2.2
Banking panic—fall
1930 36.1 40.2 40.7 40.4 36.0 43.1
Banking panic—fall
1931 12.4 11.0 10.9 10.7 8.9 11.3
Banking panic—winter
1933 2.5 3.9 3.8 3.8 2.4 1.5
nificant, indicating that during the crisis, banks in the 6th District failed
at lower rates than banks in the 8th District. We cannot reject the null
hypothesis that the other coefficients equal zero.
Table 6 reveals the magnitudes of the coefficients. Column 1 indicates
that the crisis in the fall of 1930 raised bank failure rates substantially.
The marginal effects can be stated as changes in cumulative hazard rates
(a metric readily comparable to that of the graphs in the previous sec-
tion). The regression coefficients, the parametric assumptions concern-
ing the survival function, and the data can be combined to estimate the
probability of distress for each bank for each day of a crisis period. For
the crisis during the fall of 1930, the mean estimate is 1.74 failures per
day per thousand banks. A counterfactual—what would the hazard rate
have been during the panic in the absence of the Atlanta Fed’s inter-
vention—can be estimated by setting the indicator variables for the
panic/Federal Reserve District 6 interaction equal to zero and redoing
the calculation. The mean estimate for the no-intervention counterfac-
tual is 2.43 per thousand. Another counterfactual—what would the haz-
ard rate have been in the absence of the panic—can be estimated by
setting the indicator variables for panic and panic/Federal Reserve Dis-
trict 6 interaction equal to zero. The mean estimate for the no-panic
counterfactual is 0.11 per thousand. The difference between the coun-
terfactuals indicates the impact of the panic. The difference is 2.32 per
thousand. Compounding that figure over the 155 days of the crisis re-
veals that the panic increased the cumulative hazard for the average
bank by 36.0 percent. Similar calculations indicate the effect of the
Atlanta Fed’s expansionary policy.
Columns 2–6 in tables 5 and 6 strengthen this supposition. Column
2 adds to the explanatory variables a vector of bank characteristics. The
characteristics include the percentage of total assets composed of cash,

exchanges with banks, and marketable securities (assets % cash); net
worth as a share of total assets (net worth/total assets); the natural log
of total assets; the percentage of liquid assets held as cash or deposits
in banks; the percentage of noncash assets invested in real estate; the
number of years that the bank had been in operation; whether the bank
possessed a state charter; and the share of deposits at all banks in the
county held at the bank under observation. We report coefficients for
the first two variables. The remainder can be found in the online version.
In all our specifications, we correct standard errors for heterogeneity
using the Huber-White sandwich method with error terms clustered on
individual banks.
Column 3 adds to the regression the characteristics of the counties
within which each bank operated. Column 3 adds the ratio of aggregate
farm debt to farm value, a measure of recent changes in farm values,
the percentage of farm acres in pasture or fallow, and the percentage
of farms under 100 acres. Rather than accounting for county charac-
teristics by choosing a subset of the numerous available variables, col-
umns 4 and 5 add to the regression the five principal components (as
identified by the Kaiser criterion) of 22 county-level variables that appear
particularly pertinent to Mississippi. Employing the principal compo-
nents changes neither the signs nor the significance levels of variables
concerning the banking crises and Federal Reserve policy regimes and
changes their magnitudes only slightly.
Column 5 performs a robustness check. It limits the sample to banks
located in counties within 1 degree latitude of the Federal Reserve
district border. This group of banks operated in a narrow geographic
range and experienced similar economic shocks. Limiting the sample
to this group has little influence on the magnitudes, signs, and signif-
icance levels of the coefficients.
Column 6 estimates the canonical Calomiris and Mason version of
the model. We format our data as in their 2003 essay, employing nearly
identical bank, county, state, and national data, and replicate their re-
sult. The regression does an excellent job of predicting the longevity
of individual institutions. Fundamentals are highly correlated with bank
distress. However, our version of the model includes indicators for Fed-
eral Reserve policy regimes. The coefficients on these indicators dem-
onstrate that the Federal Reserve could lower bank failure rates by acting
as a lender of last resort during banking panics.
V. Robustness
Our conclusion remains robust to a wide variety of alterations in our
econometric framework. Parametric models employing different para-
metric assumptions, explanatory variables, and corrections for hetero-

geneity and serial correlation yield identical qualitative and similar quan-
titative results. Nonparametric analysis demonstrates that our results do
not depend on particular mathematical and statistical assumptions. Both
types of analysis demonstrate that differences in the observed charac-
teristics of banks and the environments in which they operated do not
drive our results. Our results arise from patterns in the raw data apparent
from whatever perspective one views the evidence. Moreover, since all
our parametric models include corrections for unobserved heteroge-
neity and selection on unobserved characteristics and since our non-
parametric models examine subpopulations defined by factors likely
correlated with unobserved and/or selected characteristics, unobserved
differences among banks are unlikely explanations for the patterns that
appear in the data.
Several crucial issues, however, cannot be addressed statistically. Could
some unmeasured fundamental shock explain differences between the
6th and 8th Districts during the post-Caldwell crisis? To be consistent
with the evidence, the shock would have to be one that raised failure
rates in the 8th District relative to the 6th District during the period
beginning December 19, 1930, and ending March 2, 1931, but neither
before nor after, and the shock would have to be one that affected the
districts uniformly and retained its punch right up to the border but
did not spill over into the adjoining district. The shock could not be
one that we have controlled for both parametrically and nonparamet-
rically. Such shocks include anything correlated with the characteristics
of banks—such as size, age, services, financial characteristics, or Federal
Reserve membership—or the economic or demographic characteristics
of the towns or counties in which banks operated—such as population
density, number of manufacturing establishments, and cotton cultiva-
tion. These facts seem to rule out all possible climatic, cultural, agri-
cultural, and industrial shocks, all of which would seem to be correlated
with our controls or to operate on time horizons longer than 10 weeks.
Could the confounding factor be financial links to the Caldwell con-
glomerate or geographic proximity to the locus of the post-Caldwell
panic? The evidence indicates otherwise. Consider the case of financial
linkages. One of our sources, Rand McNally, lists the correspondents
for all banks in Mississippi. Another source, the St. 6386 reports in
archives of the Board of Governors, indicates whether a correspondent’s
closure caused the suspension of a client. These sources show that no
links existed between banks in Mississippi and the Caldwell conglom-
erate or its subsidiaries. This evidence of absence confirms statements
made by Mississippi’s superintendent of banks, J. S. Love, during a press
conference on November 22, 1930: “Our [Mississippi’s] banks are free
from outside allied connections. There does not exist in this state any
group or chain banking system. . . . [We] see no cause for alarm”

(Vicksburg Sunday Post-Herald, November 23, 1930, 11; Meridian Star, No-
vember 23, 1930, 1–2). Finally, including the matrix of correspondent
linkages on the right-hand side of our regressions alters neither the
signs nor the significance levels of our coefficients.
Now, consider the case of geographic proximity. In Mississippi, bank
runs began 6 weeks after Caldwell’s demise and 3 12 weeks after the last
bank in another state failed as a result of correspondent links to the
Caldwell conglomerate. Runs began in the center of Mississippi, not in
close proximity to borders of states engulfed by Caldwell’s collapse. In
addition, although the eastern half of Mississippi lay closer to Nashville,
which contained Caldwell’s headquarters, the bulk of Caldwell’s finan-
cial operations, and its largest banking affiliate, the pattern of failures
did not differ in the eastern and western halves of Mississippi or on the
basis of distance from Nashville.
Could the confounding factor be some difference in policy between
the districts other than discount lending? One potential candidate is
open-market purchases. But for both districts, discount lending far ex-
ceeded open-market purchases. Between September 7 and December
28, 1931, for example, the quantity of U.S. government securities pos-
sessed by the 8th District did not change at all and the quantity possessed
by the 6th District increased by only $4,000. At the same time, the
quantity of discounts on the balance sheet of the two districts fell by
roughly $4,000,000 and $7,000,000, respectively. Moreover, when the
districts purchased eligible paper and government securities, they did
so as an adjunct to discount lending, in order to provide favorable terms,
expedite the process of converting assets to cash, and quickly provide
liquidity to specific banks. The quantities of assets that the districts
purchased were never large enough to influence macroeconomic ag-
gregates such as the deflation or risk-free interest rate. Such macroeco-
nomic aggregates neither differed between districts nor varied substan-
tially during the event. So, they cannot explain interdistrict differences
in bank survival rates.
Another potential candidate is bank standards and supervision. But
nine out of 10 banks in Mississippi were state-chartered institutions.
Mississippi applied identical standards and examination procedures in
the northern and southern sections of the state. The Biennial Report of
Mississippi’s Banking Department, which lists the names of the exam-
iners and the institutions that they examined, indicates that examiners
rotated among institutions throughout the state. Mississippi’s state bank-
ing codes required that banks be examined “at least twice each year at
irregular intervals without prior notice, and with no bank to be ex-
amined by the same examiner twice in succession” (Warburton 1955,
15). So, interdistrict differences in regulations and examination pro-
cedures did not exist.

Another potential candidate is bailouts and subsidies. But, neither
the state government nor the Federal Reserve district banks provided
such assistance, and no banks in Mississippi received assistance from the
Reconstruction Finance Corporation until 1932. Affirmative evidence
of the absence of subsidies and bailouts exists. The Federal Reserve
Board’s form St. 6386c contains a section that describes changes in
financial structure and assistance received toward reopening. These
forms indicate that no banks that reopened following the post-Caldwell
panic received subsidies or changed their financial structures in any way.
The Federal Reserve Board’s form St. 6386b, which records bank sus-
pensions, contains a section describing borrowings from the Federal
Reserve, the RFC, and similar institutions. These forms indicate that
none of the state banks that closed their doors (temporarily or per-
manently) during the post-Caldwell panic held such loans.
Could the confounding factor be some other unmeasured shock or
policy? To answer that question, we (i) scrutinized seven newspapers
(named in Sec. II) for the months of September 1930 through March
1931, (ii) read the annual reports of Mississippi’s banking commissioner
for the years 1928–37, (iii) read the annual reports and monthly bul-
letins of the Federal Reserve 6th and 8th Districts, and (iv) scrutinized
records of bank failures collected by the Federal Reserve Board. All
these sources described the epidemic of bank runs that occurred in
Mississippi at that time. None described a shock to the economy or
differences in policies (other than discount lending) that might have
caused more banks to fail in the northern than in the southern half of
the state. It seems unlikely that such a large number of observers, with
the knowledge needed to detect such an unusual and sizable shock and
with the ability and incentive to report it, would have failed to report
such an event, if it had occurred.
What about selection? Selection could have operated through several
channels including the opening of new banks, closing of old banks, and
migration of banks between districts. In each of these cases, banks likely
to benefit from a supportive discount window because they possessed
less liquid portfolios would grow as a proportion of the banks in the
6th District, whereas banks that did not perceive the need for assistance
during panics because they possessed more liquid portfolios would grow
as a percentage of the banks in the 8th District. This process of selection
would concentrate banks susceptible to panics in the 6th District. The
concentration could cause the efficacy of monetary intervention to be
understated, since the treatment group consisted disproportionately of
vulnerable institutions.
The extant evidence, however, allays such concerns. First, when given
the option of reporting to the Federal Reserve district bank supervising
the other portion of the state at the founding of the Fed in 1913, none

of Mississippi’s banks chose to do so. Second, statistical tests cannot
reject null hypotheses that bank survival, failure, and establishment rates
in the 6th District equaled those in the 8th District between 1916 and
1928, the predepression years for which we have data. Third, statistical
tests cannot reject the null hypotheses that in 1929, banks possessed
similar asset portfolios and similar numbers of correspondents in the
6th and 8th Districts.
Selection might have operated through other channels. Managers and
depositors are also mobile. Careful managers who worried about panics,
foresaw the need for liquidity, and believed that the 6th District would
provide more liquidity than the 8th might have migrated to the 6th
District. They may also have been better judges of credit, been more
efficient, and kept more cash on hand. Depositors might also have
anticipated benefits from the 6th District’s policies and have shifted
funds toward the district that promised to provide liquidity. Either re-
action might have made banks in the 6th District stronger than those
in the 8th District. In this case, the efficacy of monetary intervention
would be overstated.
The extant evidence, once again, allays such concerns. First, a sample
of bank presidents, vice presidents, managers, and cashiers drawn ran-
domly from 50 banks (approximately one-sixth of those in Mississippi)
for the years 1915, 1925, 1929, and 1930 shows no shifts of management
between the 6th and 8th Districts. Second, Warburton’s study of banking
in Mississippi found no significant shifts in distribution of deposits from
1915 through 1929. Throughout this period, roughly the same per-
centage of deposits was held by state banks, by failed banks, and by the
five largest banks. The five largest institutions, for example, held 25.4
percent of the deposits in 1915 and 24.3 percent in 1929 (Warburton
1955, 31–36). Third, data from banks near the Federal Reserve district
boundary show no change in the quantity of deposits at banks in the
6th District relative to the 8th District between July 1929, just prior to
the suspension of deposit insurance, and July 1930, just after Mississippi
discontinued its deposit insurance system, a point in time when the
danger of bank runs, and thus liquidity assistance, increased suddenly
and substantially. Fourth, deposits did not flow from the St. Louis to
the Atlanta District in the wake of the post-Caldwell panic. In fact, banks
in the 8th District near the border lost fewer deposits than banks farther
from the border, all else held equal, and for banks operating near the
border, average deposits at banks that remained in business in the 8th
District rose relative to average deposits at banks that remained in busi-
ness in the 6th District. These patterns are the opposite of what one
would expect, if proximity to the border induced the flight of deposits.
Several factors explain the absence of selection. First, the public may
not have been aware of policy differences between the 6th and 8th

Districts. The St. Louis Fed did not advertise its opposition to interven-
tion or the way in which it operated the discount window. Moreover,
since no panics occurred in the 8th District between the founding of
the Fed and the Great Depression, the St. Louis Fed never demonstrated
the actions that it would take in such an event.
Second, depositors and bankers may have underestimated the like-
lihood and severity of a potential banking panic, because severe banking
panics had not occurred in Mississippi or on a national scale for a
generation, and because during the Roaring Twenties, few people ex-
pected the onset of a catastrophic contraction. Thus, depositors and
bankers may not have anticipated the need for a lender of last resort.
Third, the public may not have anticipated beneficial effects from
monetary intervention. Debates over the effectiveness of the policy have
raged for at least two centuries, dating back to Hume’s writings on the
topic and continuing vigorously today. The benefits of the approach
were disputed during the 1920s. Leading academics, bankers, business-
men, and policy makers, including much of the leadership of the Federal
Reserve System, believed that discount lending would exacerbate, rather
than alleviate, the situation. It is unclear what depositors believed about
the topic or if they had any beliefs at all.
Fourth, from 1914 until 1930, Mississippi operated a statewide deposit
insurance system. Its existence may have rendered the Atlanta Fed’s
assistance superfluous and may also have reduced depositors’ attention
to the issue. Mississippi’s superintendent of banking believed this to be
the case. He repeatedly wrote that deposit insurance discouraged de-
positor monitoring and therefore encouraged mismanagement. For this
reason, Mississippi discontinued its deposit insurance system on March
13, 1930 (Mississippi Banking Department Biennial Report 1929, 4–9;
1931, 4–5).
Fifth, even if the public had possessed perfect foresight, shifting from
one district to another may not have been in their best interest. Bank
managers’ ability to attract deposits depended on their standing within
their community and their reputation for honesty, reliability, and fi-
nancial acumen. Their ability to earn profits depended on personal
knowledge of individuals and businesses and their success at using that
knowledge to assess the risks and returns of extending credit. Moving
to a new location meant abandoning the informational and reputational
advantages that enabled individuals to operate banks profitably. Shifting
deposits to distant towns also entailed disadvantages. Holding deposits
at a distance made it more difficult to monitor the health of one’s bank
and made it more difficult to withdraw funds during a panic, when
individuals at the head of the line received the full value of their de-
posits, and those at the end of the line lost a large portion of their life
savings.

TABLE 7
Asset Quality at Suspended Banks in Mississippi, January 1929
through March 1933
Good Problematic
Number % Number %
6th Atlanta:
1. Panic 3 25.0 9 75.0
2. Nonpanic 2 8.0 23 92.0
8th St. Louis:
3. Panic 37 55.2 30 44.8
4. Nonpanic 12 38.7 19 61.3
Source.—National Archives and Records Administration, Record Group 82. See Richardson (2006) for
details.
Note.—Rows 1 and 3 present figures for all banks suspending during October, November, and December
1930 and January, February, and March 1931. Rows 2 and 4 present figures for banks suspending operations
in all other months from January 1929 through March 1933. The rows sum to 100 percent. Columns indicate
the percentage of suspended banks in each district in each period whose assets were judged by examiners
to be good and, thus, to have been neither a primary nor a contributing cause of the suspension, and problematic
(i.e., either slow, doubtful, or worthless) and to have been either a primary or a contributing cause of the
suspension.
A final type of evidence completes the case. Three sources report the
quality of assets at failed banks. First, the Federal Reserve Board St. 6386
forms indicate examiners’ ex ante (i.e., before the suspension) assess-
ment of the quality of assets at suspended banks. Examiners reported
the quality of assets at banks to be good and neither to have been a
primary nor a contributing cause of the suspension, or problematic (i.e.,
either slow, doubtful, or worthless) and to have been either a primary
or contributing cause of the suspension. Table 7 presents this infor-
mation. It shows that the quality of assets at institutions that suspended
operations in the 8th District was better than the quality in the 6th
District. During the post-Caldwell panic, the majority of the banks that
suspended operations in the 8th District had portfolios consisting pre-
dominantly of good assets.
Second, Warburton’s study, “Deposit Guaranty in Mississippi,” pro-
vides evidence on recoveries from the assets of failed banks (Warburton
1955, 41–51, tables 11–13). From 1916 to March 1930, when Mississippi
guaranteed bank deposits, recoveries averaged just over 51.5 percent
(i.e., on average, assets with a book value of $100 yielded $51.50). Re-
coveries from the assets of banks that failed during the post-Caldwell
panic averaged 70.4 percent. Third, the 1929, 1930, and 1931 Biennial
Reports of Mississippi’s Banking Department (tables F and G) record
information on recoveries from banks in liquidation. For 39 banks that
failed during the post-Caldwell panic, data exist on (a) recoveries from
the initial sale of assets shortly following suspension (these sales were
supposed to be of assets that yielded nearly book value or better), (b)
the initial estimate of the value of the remaining assets, and (c) eventual
recoveries from sales of the remaining assets during the years 1931–33.

After the initial liquidation, examiners estimated the total value of the
assets remaining from banks in the 6th District to be $1,022,025. By the
end of 1933, recoveries from the sale of those assets equaled $1,014,735
(i.e., 99.3 percent of their estimated value). After the initial liquidation,
examiners estimated the total value of the assets remaining from banks
in the 8th District to be $2,738,760. By the end of 1933, recoveries from
the sale of those assets equaled $2,179,231 (i.e., 79.6 percent of their
estimated value).
Together, the sources illuminate important patterns over time and
across districts. Over time, banks that failed during the panic of 1930,
when liquidity forced banks with cash flow problems to shut their doors,
were healthier than banks that failed at other times, when healthy in-
stitutions had access to liquidity. Across districts, banks that failed in
the 8th District, where the Federal Reserved did not act as a lender of
last resort, were healthier than banks that failed in the 6th District,
where the Federal Reserve strove to expand the supply of credit. After-
ward, in the 8th District (where more and healthier banks failed), re-
ceivers sold substantial quantities of bank assets (primarily loans to local
businesses, consumers, and farmers), and the values of those assets de-
clined precipitously. In the 6th District (where prompt intervention by
the Federal Reserve enabled many banks to weather the storm), receivers
put fewer assets on the market, and the values of those assets declined
to a lesser degree.
VI. Impact on Commercial Activity

Did bank failures harm the broader economy? Did the destruction of
financial institutions impair industry and commerce? Would saving
banks have had beneficial economic effects?
Patterns of credit and commerce over time and across districts suggest
that bank failures did, in fact, restrict wholesale trade. These patterns
are illustrated by figure 6 and table 8, which show that after the banking
panics, credit and commerce contracted more in the 8th District’s por-
tion of Mississippi than in the 6th District’s portion. Figure 6A plots
total deposits at banks in the 6th and 8th Districts’ portions of Mississippi
from 1930 through 1933. Deposits declined less in the 6th and more
in the 8th District. Controlling for prepanic trends in deposits magnifies
the difference. Figure 6B plots total loans and discounts, which is a
reasonable proxy for total lending to local businesses. Lending declined
less in the 6th and more in the 8th District. Controlling for prepanic
trends magnifies this difference. Table 8 reports patterns of wholesale
trade, an economic sector that depended on banks to finance ongoing
operations and short-term working capital. Between 1929 and 1933, the
number of firms in the 8th District’s portion of Mississippi fell by 34.7

Fig. 6.—Comparing consequences of the banking panics in the 6th and 8th Districts.
A, Total deposits as a percentage of total deposits in June 1930. B, Total loans and discounts
as a percentage of the total in June 1930.
percent, a failure rate nearly twice that of firms in the 6th District.
Wholesale transactions also fell farther in the 8th than in the 6th District.
Variation in the reason for and location of bank failures enables us
to draw clear conclusions concerning the contraction of credit and
decline in commercial activity. The regressions reported in table 9 ex-
ploit this variation. The dependent variable is the decline in wholesale
transactions (measured as 1929 real dollars) in each county from 1929
to 1933 (measured by summing net sales for all wholesalers in each
county during the year 1929 and subtracting the same sum for the year
1933). Baseline specifications appear in columns 1 and 2. Column 1’s
sole explanatory variable is the decline in total loans outstanding from
1929 to 1933 (measured by summing loans and discounts on the balance

TABLE 8
Decline in Wholesale Trade
Federal Reserve District

6th Atlanta 8th St. Louis
Wholesale firms:
Number in 1929 783 930
Number in 1933 641 607
D% ⫺18.1 ⫺34.7
Net sales:
$1,000s in 1929 140,776 245,486
$1,000s in 1933 59,513 83,727
D% ⫺57.7 ⫺65.9
Source.—Census of American Business, 1929 and 1933.
sheets of all banks in each county in operation on July 1 of each year).

Note that the dependent variable measures the change in a flow, whole-
sale transactions summed up over the entire year minus wholesale trans-
actions summed over a later year. The explanatory variable measures
the change in a stock, total lending on the balance sheets of banks on
the first business day in July minus the same figure at a later date. The
link between the stock and flow is the average maturity of short-term
TABLE 9
County-Level Regressions of Commercial Activity on Credit Contraction
Dependent Variable: Decline in Net Wholesale Transactions from 1929 to 1933
(1) (2) (3) (4) (5) (6)

Decline in loans, 1929–33 1.55* 1.21*
(.54) (.60)
Decline in loans due to all 2.54*
bank liquidations, 1929–33 (.77)
Decline in loans due to bank 4.00* 3.98* 5.37*
liquidations during panic of (1.00) (.96) (.68)
1930–31
Decline in loans due to bank .87 .90 ⫺.88
liquidations outside of panic (.63) (.67) (1.05)
period
Decline in loans at surviving ⫺.52 ⫺.28 ⫺.30 ⫺.45
banks (.59) (.41) (.42) (.40)
Loans at banks suspending .155 ⫺.96**
temporarily during panic of (.51) (.53)
1930–31
Constant (1,000s) ⫺681** ⫺1,009* ⫺952* ⫺949* ⫺907* ⫺1,498*
(376) (561) (451) (400) (453) (536)
Vector of county characteristics Yes Yes Yes Yes Yes
Sample limited to counties
near border Yes
Observations 82 82 82 82 82 39
R2 .36 .64 .77 .80 .81 .93
Note.—Standard errors are in parentheses.
* Significant at the 5 percent level.
** Significant at the 10 percent level.

financial instruments used to finance wholesale transactions, such as
commercial paper and bankers’ acceptances. That average was just over
90 days during the period under consideration. This maturity meant
that each dollar a bank invested in short-term commercial paper turned
over four times each year and financed $4 of wholesale trade. The
coefficient of 1.55 on the variable decline in loans, 1929–33, therefore,
has a sensible magnitude. It suggests that when banks lost $1 in lending
capacity, they reduced lending to wholesalers by 39 cents (i.e., 1.55
divided by 4) and reduced lending to others by 61 cents. Column 2
adds a vector of county characteristics (identical to the vector of prin-
cipal components used in table 5) to the set of explanatory variables.
The correlation between wholesale activity and bank lending remains
statistically significant and retains a reasonable magnitude. Column 3
separates the decline in lending into two components: the decline in
lending due to bank failures (measured as the dollar value of lending
trapped in banks sent to receivership) and the decline in lending at
banks remaining in operation. The first component remains significant
in the statistical and practical sense. The second component’s coefficient
is not statistically significant and has a magnitude much closer to zero.
The decline of lending capacity at banks remaining in operation, in
other words, appears uncorrelated with the decline in wholesale activity.
Column 4 subdivides the decline in lending into three components: the
decline in lending due to bank liquidations during the panic of 1930–
31, the decline due to bank liquidations in all other periods, and the
decline at banks remaining in operation. The null hypothesis that the
coefficient equals zero can be rejected only for the first component. In
other words, the loss of lending capacity following the panic of 1930–
31 was significantly and substantively correlated with declines in whole-
sale activity. The loss of lending capacity for other reasons was not.
Column 5 checks the robustness of that result by adding an explanatory
variable measuring lending at banks that suspended operations tem-
porarily during the banking panic of 1930–31. This variable, which con-
trols for the severity of panics orthogonal to bank liquidations, is
insignificant.
The correlation between bank failures and business activity appears
to be robust. Column 6 shows that the pattern holds even after restricting
the sample to counties lying within 1 degree of the Federal Reserve
district border. Additional regressions using the number of wholesale
firms, the number of retail firms, the volume of retail trade, and tax
revenues yield similar signs and significance levels.
The magnitudes of the coefficients in table 9 have substantive inter-
pretations. During the 1930s, bank portfolios contained large quantities
of short-term commercial paper with maturities averaging approximately
3 months. Thus, $1 of bank lending power financed approximately $4

in wholesale transactions during a year. Extending short-term commer-
cial credit was a principal function of commercial banks. Facilitating
the operation of this market was a principal mission of the Federal
Reserve. Until 1932, short-term commercial paper was the only asset
deemed eligible for rediscount at the Federal Reserve discount window.
If the disruption of the banking system influenced real economic activity,
it should certainly have affected short-term financing for wholesale
firms. This appears to be the case. Specifications 4 and 5 indicate that
by locking $1 of lending power in banks thrown into receivership, the
banking panic of 1930 took $1 of credit from the market of commercial
paper, reducing wholesale transactions by $4 annually.
The two stages of our analysis enable us to draw causal conclusions
linking monetary policy to bank behavior to economic outcomes. The
first stage showed that discount lending helped banks survive financial
panics. The second stage showed that bank failures during the panic
reduced lending available for wholesale activity. Bank distress at other
times (e.g., during the fall of 1931) and for other reasons (e.g., loan
defaults) shows little (or no) correlation with the commercial contrac-
tion.
VII. Discussion
The multiple sources and methods employed in the previous sections
tell a consistent tale. During the banking panic that began in December
1930, banks failed at lower rates in the 6th Federal Reserve District,
where the Atlanta Fed injected liquidity into the banking system, than
in the 8th Federal Reserve District, where the St. Louis Fed followed
the doctrine of real bills. The St. Louis Fed could have followed the
same policy as the Atlanta Fed, and if it had, bank failure rates would
have been lower, commercial lending would have remained higher, and
the contraction would not have been as severe.
The quasi-experimental structure of our study, which frees our esti-
mates from difficulties of inference that typically trouble studies of firms
in complex, changing, and endogenous economic environments,
strengthens our conclusion. The limitations of our analysis are the same
as those for any study of this type. While our methods generate a precise
and powerful result, they do so for a particular point in time and space:
an agricultural state during the initial banking panic of the 1930s. The
generalizability of our result depends on the representativeness of the
place and period under study. On this dimension, our study stands on
strong ground.
Mississippi’s banks were representative of the segments of the financial
system that bore the brunt of the contraction. Mississippi was an agri-
cultural state suffering from droughts, falling commodity prices, and

the broad economic downturn that followed the stock market crash.
Unit banks predominated. Most banks possessed state charters. Similar
conditions existed in the regions of the nation and segments of financial
industry that suffered the bulk of all bank failures.
Mississippi’s banks were also representative of the segments of the
financial system crucial for understanding links between banking panics,
monetary policies, and the real economy. Most banks in Mississippi were
medium- to small-sized state-chartered institutions. Their customers
tended to be individuals, farmers, and businesses lacking access to equity
markets and other nonbank sources of credit. Their managers possessed
information about local borrowers that was lost when they ceased op-
erations. Bernanke (1983) identifies the destruction of this information
and the resulting disintermediation as one of the channels by which
financial crises exacerbated the Great Depression. In addition, the
medium- to small-sized banks in Mississippi were typical of the institu-
tions that bore the brunt of the deposit losses during the early years of
the depression, as depositors shifted funds toward larger member banks
that were less likely to fail or removed funds entirely from the depository
system. Medium- to small-sized state banks were also the institutions that
accumulated the largest excess reserves. Monetarists identify declines of
the deposit-currency and deposit-reserve ratios as principal factors be-
hind the collapse of the money supply and aggregate economy between
1931 and 1933 (Friedman and Schwartz 1963). Moreover, the collapse
of state banking systems that began in the fall of 1930 received prom-
inent media coverage. The widespread reporting of the bank failures—
including incessant coverage of defalcations, indictments, and suicides
of bankers—must have generated fear and uncertainty among consum-
ers and businessmen. Romer (1993) among others identifies uncertainty
and expectations as mechanisms by which financial crises deepened the
depression. Thus, the banking situation in Mississippi during the 1930s
reflects the three primary channels—money, intermediation, and ex-
pectations—by which bank failures influenced real economic activity.
Was concerted action by the Federal Reserve feasible? The weight of
the evidence suggests that the Federal Reserve System could have done
more to combat the initial wave of banking panics in the fall of 1930.
At that time, the Federal Reserve knew the theoretical justification for
intervening to halt banking panics. The Federal Reserve also had the
ability to act as a lender of last resort. Gold stocks were large. Gold was
flowing into the country. Credit could have been extended to banks
without endangering the exchange rate regime.
The policies followed by the Atlanta Fed required the commitment
of few resources. The extra manpower required to extend emergency
loans, the forms that had to be filled out, and the fuel required to move
cash from the Fed to commercial banks were trivial. The financial costs

were also limited. The Federal Reserve Bank of Atlanta was not saddled
with large liabilities like those the government incurred during the
savings and loan crisis of the 1980s. In contrast, the Atlanta Fed profited
by extending emergency credit, because the loans were repaid. Atlanta’s
credibility may have been one reason for the low cost of its policies. In
expectational panics of the Diamond-Dybvig type, a lender of last resort
that credibly commits to fulfilling its mission may expend fewer re-
sources than a central bank that makes a belated and halfhearted at-
tempt to halt a panic. Atlanta’s experience demonstrates, in other words,
that the costs of action are sometimes less than the costs of inaction.
Evidence that the Atlanta Fed’s policies influenced broader business
conditions comes from several sources in addition to the regressions
presented in the previous section. Economic historians have shown that
the depression followed a unique course in the 6th Federal Reserve
District. The 6th District experienced a contraction during 1929 and
1930 as sharp and severe as the hardest-hit Federal Reserve districts,
but the 6th District’s recovery began earlier and progressed swifter than
anywhere else in the United States. The 6th District’s recovery began
during the first quarter of 1931, which was the quarter that the Atlanta
Fed embarked on its efforts to extend discount loans to banks and
expand the supply of credit. During that quarter, the contraction ac-
celerated in all other districts, and by 1933, the 6th District’s economy
was the healthiest in the nation. This pattern appears in data on em-
ployment and unemployment (Wallis 1989) and indices of industrial
production (Rosenbloom and Sundstrom 1999). The pattern also ap-
pears in data on banks. In the 6th District, depositors kept more of
their savings in the financial system, bankers held lower excess reserves
and made more loans, and businesses borrowed more money. Scholars
refer to the pattern as the Southern Paradox, since they have failed to
find any features of the southern economy that can explain the South’s
sudden, singular recovery (Margo 1993).
Several strands of literature suggest that stopping the banking panics
would have moderated the contraction for the nation as a whole. First,
much of the scholarship on the Great Depression, including Friedman
and Schwartz (1963) and Bernanke (1983), sees banking panics as a
principle propagator of the contraction and a key to understanding the
downturn’s depth and length. Second, many macroeconomic models
highlight a connection between the onset of banking panics in the fall
of 1930 and the depression’s acceleration at that time. For example,
Cecchetti and Karras (1994) find that “there is an aggregate supply
collapse that coincides with the onset of severe bank panics . . . sug-
gesting an association between [the supply shocks] and the credit chan-
nels emphasized by Bernanke” (80–81, 99–100). Christiano, Motto, and
Rostagno (2003) find that the flight from deposits to currency during

the year following Caldwell’s collapse (and the consequent accelerator
effects, debt deflation, and credit crunch) explains the severity of the
contraction during the years 1931–33. Third, monetarists such as Fried-
man and Schwartz (1963) argue that the onset of banking panics in the
fall of 1930 led to the collapse of the monetary system. The initial panic
marks the point at which the money multiplier, the deposit-currency
ratio, the deposit-reserve ratio, and all measures of the money supply
plummeted. Their decline from November 1930 to March 1933 was the
most rapid and prolonged in American economic history. Their decline
lowered the price level, raised real wages and interest rates, and, through
those and other channels, reduced both aggregate supply and demand.
If those scholars are correct, then the evidence presented in this essay
indicates that the Federal Reserve System missed an opportunity to take
inexpensive actions that would have stemmed the initial wave of banking
panics and altered the course of the contraction. The broader impli-
cations of this finding remain to be determined. Would mitigating the
initial wave of panics have prevented the panics that came later? Would
mitigating panics have prevented the debilitating deflation and collapse
of intermediation that dragged the U.S. economy ever deeper into de-
pression? What were the relative strengths of the money and credit
channels for the transmission of monetary policy? All these inquiries
remain open questions. The evidence presented in the preceding par-
agraphs is only suggestive. However, we believe the approach that we
pioneer—applying quasi-experimental methods to panels of data on
banks and businesses exposed to different monetary regimes along Fed-
eral Reserve district borders—can be extended to answer these and
other questions concerning monetary policy, financial intermediation,
and the causes, consequences, and possibilities of preventing Great
Depressions.
References
Bernanke, Benjamin S. 1983. “Nonmonetary Effects of the Financial Crisis in
the Propagation of the Great Depression.” A.E.R. 73 (June): 257–76.
Calomiris, Charles W., and Joseph R. Mason. 2003. “Fundamentals, Panics, and
Bank Distress during the Depression.” A.E.R. 93 (December): 1615–46.
Cecchetti, Stephen G., and Georgios Karras. 1994. “Sources of Output Fluctu-
ations during the Interwar Period: Further Evidence on the Causes of the
Great Depression.” Rev. Econ. and Statis. 76 (January): 80–102
Chandler, Lester Vernon. 1971. American Monetary Policy, 1928–1941. New York:
Harper and Row.
Christiano, Lawrence J., Roberto Motto, and Massimo Rostagno. 2003. “The
Great Depression and the Friedman-Schwartz Hypothesis.” J. Money, Credit,
and Banking 35, no. 6, pt. 2 (December): 1119–97.
Diamond, Douglas W., and Philip H. Dybvig. 1983. “Bank Runs, Deposit Insur-
ance, and Liquidity.” J.P.E. 91 (June): 401–19.

Diamond, Douglas W., and Raghuram Rajan. 2005. “Liquidity Shortages and
Banking Crises.” J. Finance 60 (April): 615–47.
———. 2006. “Money in a Theory of Banking.” A.E.R. 96 (March): 30–53.
Eichengreen, Barry. 1992. Golden Fetters. New York: Oxford Univ. Press.
Friedman, Milton, and Anna J. Schwartz. 1963. A Monetary History of the United
States, 1867–1960. Princeton, NJ: Princeton Univ. Press.
Gamble, Richard. 1989. A History of the Federal Reserve Bank of Atlanta, 1914–1989.
Atlanta: Fed. Reserve Bank Atlanta.
Margo, Robert A. 1993. “Employment and Unemployment during the 1930s.”
J. Econ. Perspectives 7 (2): 41–59.
Meltzer, Allan H. 2003. A History of the Federal Reserve. Vol. 1, 1913–1951. Chicago:
Univ. Chicago Press.
Richardson, Gary. 2006. “The Records of the Federal Reserve Board of Governors
in the National Archives of the United States.” Financial Hist. Rev. 13 (April):
123–34.
———. 2007a. “Bank Distress during the Great Depression: The Illiquidity-
Insolvency Debate Revisited.” Explorations Econ. Hist. 44 (October): 586–607.
———. 2007b. “The Check Is in the Mail: Correspondent Clearing and the
Banking Panics of the Great Depression.” J. Econ. Hist. 67 (September): 643–
71.
———. 2008. “Quarterly Data on the Categories and Causes of Bank Distress
during the Great Depression.” Res. Econ. Hist. 25 (January): 37–115.
Richardson, Gary, and William Troost. 2006. “Monetary Intervention Mitigated
Banking Panics during the Great Depression: Quasi-Experimental Evidence
from the Federal Reserve District Border in Mississippi, 1929 to 1933.” Working
Paper no. w12591 (October), NBER, Cambridge, MA.
Romer, Christina. 1993. “The Nation in Depression.” J. Econ. Perspectives 7
(Spring): 19–39.
Romer, Christina, and David Romer. 1989. “Does Monetary Policy Matter? A
New Test in the Spirit of Friedman and Schwartz.” In NBER Macroeconomics
Annual, edited by Olivier Blanchard and Stanley Fischer. Cambridge, MA: MIT
Press.
Rosenbloom, Joshua, and William Sundstrom. 1999. “The Sources of Regional
Variation in the Severity of the Great Depression: Evidence from U.S. Man-
ufacturing, 1919–1937.” J. Econ. Hist. 59 (September): 714–47.
Temin, Peter. 1976. Did Monetary Forces Cause the Great Depression? New York:
Norton.
———. 1989. Lessons from the Great Depression. Cambridge, MA: MIT Press.
U.S. Census Bureau. 1935. Census of American Business: 1933, Wholesale Distribution.
Vol. 1, Detailed Statistics for the United States, Brief Summaries by States, Cities, and
Counties. Washington, DC: U.S. Government Printing Office.
Wall, Norman. 1937. Demand Deposits of Country Banks. U.S. Department of Ag-
riculture Technical Bulletin no. 575. Washington, DC: U.S. Government Print-
ing Office.
Wallis, John J. 1989. “Employment during the Great Depression: New Data and
Hypotheses.” Explorations Econ. Hist. 26 (January): 45–72.
Warburton, Clark. 1955. “Deposit Guaranty in Mississippi.” Manuscript, Div. Res.
and Statis., Fed. Reserve Deposit Corp.
Westerfield, Ray B. 1932. “Marginal Collateral to Discounts at the Federal Reserve
Banks.” A.E.R. 22 (March): 34–55.

Wheelock, David. 1991. The Strategy and Consistency of Federal Reserve Monetary
Policy, 1924–1933. Cambridge: Cambridge Univ. Press.
Wicker, Elmus. 1996. The Banking Panics of the Great Depression. Cambridge: Cam-
bridge Univ. Press.


Reading Materials

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Reading Materials

Uploaded by

Copyright:

Available Formats

Monetary Intervention Mitigated Banking Panics

during the Great Depression: Quasi-

[ Journal of Political Economy, 2009, vol. 117, no. 6]

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

Banks failed throughout the Great Depression. Their demise contrib-

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

State Charter National Charter

II. Data Sources

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

8th Federal Reserve District

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

III. Historical Background

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

C. The Banking Crisis

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

D. The Historical Experiment

Responses to this banking panic create the historical experiment. The

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

Percentage of Banks Percentage of Banks

policies resembling those in Atlanta, also serves as a control case. If

E. Patterns Apparent in Basic Data

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

where u is the bandwidth and

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

(1) (2) (3) (4) (5) (6)

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

(1) (2) (3) (4) (5) (6)

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

VI. Impact on Commercial Activity

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

Federal Reserve District

sheets of all banks in each county in operation on July 1 of each year).

(1) (2) (3) (4) (5) (6)

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM

This content downloaded from 128.252.067.066 on July 25, 2016 22:24:36 PM